Hashes: digital fingerprints
Change one letter of a novel and its fingerprint changes completely.
Explorer · 7 min · Updated 2026-11-08
💡 A data blender
Put fruit in a blender and you get a smoothie. Same fruit, same smoothie — every time. Add one grape and the smoothie tastes completely different. And nobody can un-blend a smoothie back into fruit. A hash function is a blender for data: same input → same output; tiny change → totally different output; and you can't reverse it.
Five properties that matter
| Property | Meaning |
|---|---|
| Deterministic | Same input always gives the same hash |
| Fixed size | SHA-256 outputs 256 bits (64 hex characters), whether you hash one word or a whole movie |
| Avalanche effect | Change one bit and about half the output bits flip |
| One-way | You can't work backwards from the hash to the input |
| Collision-resistant | Finding two inputs with the same hash is practically impossible |
How blockchains use hashes
- Linking blocks: each block stores the previous block's hash. Edit an old block and its hash changes, breaking every link after it.
- Mining: Proof-of-Work means finding a block hash below a target (e.g. starting with many zeros).
- Addresses: Bitcoin addresses are derived from hashes of public keys.
- Merkle trees: thousands of transactions are summarised into one "root" hash in the block header.
Hash ≠ encryption
Encryption is reversible with the right key; hashing is not reversible at all. You hash to verify data, you encrypt to hide it.
🧮 Practical example: checking a download
Key takeaways
- Same input → same hash; tiny change → totally different hash.
- Hashes are one-way and fixed-length.
- Blocks chain together via previous-block hashes.
- Mining = searching for a hash below a target.
Education only — not financial advice.